paulmach
commented
8 years ago I need to find some time to look closely at the code but this is what i think is happening. The code tries to map "interesting" points to the closest new point. Normally there are way more new points than old points so this works okay. However, in this case it looks like the "interesting" points are in the middle of a long line that got simplified. So it maps it to the closest endpoint of that line. That is not good. It should create a new point on the line if the endpoint is too far away.
It also only maps one interesting point to one new point. So if there are 5 interesting points on a straight line that get simplified out those interesting points will get assigned to the 5 points before the line. I think.
Does that all make sense with what you're seeing?
The js code doing all this can be found here: https://github.com/strava/iD/blob/master/js/id/actions/slide.js
aawiseman
Hi, not sure this is the right place to post it, but I was looking at Trinidad and noticed what might be a bug with Slide.
It looks like a user used Slide to align a road to the Strava tracks, and it moved the nodes for where the side streets connect to the main street so they were misaligned -- not sure if that was user error or maybe something funky with Slide itself. It seems like it did the alignment of the target road successfully but moved the nodes along the line.I've attached below the image from JOSM and from regular OSM before it was fixed.
This is the changeset (I think) that caused the issue, it was awhile ago so maybe that is fixed, but it persisted until just recently when someone fixed it. https://www.openstreetmap.org/changeset/36207456 It's here specifically https://www.openstreetmap.org/#map=18/10.716196/-61.477808